The Mozambique Channel (MC) plays an important role in the exchange of water masses between Indian and Atlantic Ocean. During R/V SONNE cruise SO-183 16 lines of multibeam sonar and sub-bottom profiler data were collected in the southern MC. They show a highly variable microtopography on the seafloor. Four main microtopographic zones and several sub-zones have visually been identified. The main zones consider the overall morphology and divide the study area into regions with smooth seafloor, wavy bedforms, seamounts and islands, and the Zambezi Channel. The sub-zones take the reflection pattern and the shape, size and orientation of the bedforms into account. A smooth seafloor occurs on the Mozambican continental slope, north and south of Bassas da India, on the eastern Zambezi Channnel levee and in the Zambezi cone. Wavy bedforms cover the SW, central and NE areas. The most spectacular features are large erosional scours in the SW area. They lie in a region, where the northward flowing Antarctic Bottom Water (AABW) is deflected towards the east due to the shallowing of the MC. Farther eastwards SW-NE trending wavy bedforms are obviously aligned parallel to the deflected AABW and are therefore interpreted as contourite mounds. A W-E trending channel indicates the northernmost extension of the AABW. NW-SE oriented wavy bedforms in the west, hummocky bedforms in the east and arcuate, cross-cutting features in-between reflect a different current regime in the central area. Comparisons with LADCP measurements show, that the western part lies in the range of deep-reaching eddies, so that the wavy bedforms again seem to be contourite mounds aligned parallel to a part of the swirl. The cross-cutting features mark the eastern boundary of the eddy, where a northbound flow direction prevails. The origin of arcuate bedforms and depressions in the NE area is not clear. Deep-reaching eddies which interact with the topography of Bassas da India and the Zambezi Channel may contribute to their formation. All morphological features are draped with sediments indicating, that recent bottom-current velocities are not high enough to erode sediments. This agrees with published velocities of 0.1 m/s. Therefore, the microtopography must originate from a time, when bottom-current velocities were higher. Assuming a published sedimentation rate of 20 m/Myrs and a sediment drape of at least 60 m thickness we propose, that the microtopography developed during Pliocene times or earlier.